/*===-- Lexer.l - Scanner for llvm assembly files --------------*- C++ -*--===//
//
+// The LLVM Compiler Infrastructure
+//
+// This file was developed by the LLVM research group and is distributed under
+// the University of Illinois Open Source License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
// This file implements the flex scanner for LLVM assembly languages files.
//
//===----------------------------------------------------------------------===*/
#include <cctype>
#include <cstdlib>
+void set_scan_file(FILE * F){
+ yy_switch_to_buffer(yy_create_buffer( F, YY_BUF_SIZE ) );
+}
+void set_scan_string (const char * str) {
+ yy_scan_string (str);
+}
+
#define RET_TOK(type, Enum, sym) \
llvmAsmlval.type = Instruction::Enum; return sym
+namespace llvm {
-// TODO: All of the static identifiers are figured out by the lexer,
+// TODO: All of the static identifiers are figured out by the lexer,
// these should be hashed to reduce the lexer size
// atoull - Convert an ascii string of decimal digits into the unsigned long
-// long representation... this does not have to do input error checking,
+// long representation... this does not have to do input error checking,
// because we know that the input will be matched by a suitable regex...
//
static uint64_t atoull(const char *Buffer) {
for (char *BIn = Buffer; *BIn; ) {
if (BIn[0] == '\\' && isxdigit(BIn[1]) && isxdigit(BIn[2])) {
char Tmp = BIn[3]; BIn[3] = 0; // Terminate string
- *BOut = strtol(BIn+1, 0, 16); // Convert to number
+ *BOut = (char)strtol(BIn+1, 0, 16); // Convert to number
if (!AllowNull && !*BOut)
ThrowException("String literal cannot accept \\00 escape!");
-
+
BIn[3] = Tmp; // Restore character
BIn += 3; // Skip over handled chars
++BOut;
return BOut;
}
+} // End llvm namespace
+
+using namespace llvm;
+
#define YY_NEVER_INTERACTIVE 1
%}
/* Label identifiers end with a colon */
Label [-a-zA-Z$._0-9]+:
+QuoteLabel \"[^\"]+\":
/* Quoted names can contain any character except " and \ */
StringConstant \"[^\"]*\"
begin { return BEGINTOK; }
end { return ENDTOK; }
-true { return TRUE; }
-false { return FALSE; }
+true { return TRUETOK; }
+false { return FALSETOK; }
declare { return DECLARE; }
global { return GLOBAL; }
constant { return CONSTANT; }
-const { return CONST; }
internal { return INTERNAL; }
linkonce { return LINKONCE; }
weak { return WEAK; }
implementation { return IMPLEMENTATION; }
zeroinitializer { return ZEROINITIALIZER; }
\.\.\. { return DOTDOTDOT; }
+undef { return UNDEF; }
null { return NULL_TOK; }
to { return TO; }
-except { return EXCEPT; }
+except { RET_TOK(TermOpVal, Unwind, UNWIND); }
not { return NOT; } /* Deprecated, turned into XOR */
+tail { return TAIL; }
target { return TARGET; }
+triple { return TRIPLE; }
+deplibs { return DEPLIBS; }
endian { return ENDIAN; }
pointersize { return POINTERSIZE; }
little { return LITTLE; }
big { return BIG; }
volatile { return VOLATILE; }
+align { return ALIGN; }
+section { return SECTION; }
+module { return MODULE; }
+asm { return ASM_TOK; }
+sideeffect { return SIDEEFFECT; }
+
+cc { return CC_TOK; }
+ccc { return CCC_TOK; }
+fastcc { return FASTCC_TOK; }
+coldcc { return COLDCC_TOK; }
void { llvmAsmlval.PrimType = Type::VoidTy ; return VOID; }
bool { llvmAsmlval.PrimType = Type::BoolTy ; return BOOL; }
ulong { llvmAsmlval.PrimType = Type::ULongTy ; return ULONG; }
float { llvmAsmlval.PrimType = Type::FloatTy ; return FLOAT; }
double { llvmAsmlval.PrimType = Type::DoubleTy; return DOUBLE; }
-type { llvmAsmlval.PrimType = Type::TypeTy ; return TYPE; }
label { llvmAsmlval.PrimType = Type::LabelTy ; return LABEL; }
+type { return TYPE; }
opaque { return OPAQUE; }
add { RET_TOK(BinaryOpVal, Add, ADD); }
setle { RET_TOK(BinaryOpVal, SetLE, SETLE); }
setge { RET_TOK(BinaryOpVal, SetGE, SETGE); }
-phi { RET_TOK(OtherOpVal, PHINode, PHI); }
+phi { RET_TOK(OtherOpVal, PHI, PHI_TOK); }
call { RET_TOK(OtherOpVal, Call, CALL); }
cast { RET_TOK(OtherOpVal, Cast, CAST); }
+select { RET_TOK(OtherOpVal, Select, SELECT); }
shl { RET_TOK(OtherOpVal, Shl, SHL); }
shr { RET_TOK(OtherOpVal, Shr, SHR); }
-va_arg { return VA_ARG; /* FIXME: OBSOLETE */}
-vanext { RET_TOK(OtherOpVal, VANext, VANEXT); }
-vaarg { RET_TOK(OtherOpVal, VAArg , VAARG); }
-
+vanext { return VANEXT_old; }
+vaarg { return VAARG_old; }
+va_arg { RET_TOK(OtherOpVal, VAArg , VAARG); }
ret { RET_TOK(TermOpVal, Ret, RET); }
br { RET_TOK(TermOpVal, Br, BR); }
switch { RET_TOK(TermOpVal, Switch, SWITCH); }
invoke { RET_TOK(TermOpVal, Invoke, INVOKE); }
unwind { RET_TOK(TermOpVal, Unwind, UNWIND); }
-
+unreachable { RET_TOK(TermOpVal, Unreachable, UNREACHABLE); }
malloc { RET_TOK(MemOpVal, Malloc, MALLOC); }
alloca { RET_TOK(MemOpVal, Alloca, ALLOCA); }
store { RET_TOK(MemOpVal, Store, STORE); }
getelementptr { RET_TOK(MemOpVal, GetElementPtr, GETELEMENTPTR); }
+extractelement { RET_TOK(OtherOpVal, ExtractElement, EXTRACTELEMENT); }
+insertelement { RET_TOK(OtherOpVal, InsertElement, INSERTELEMENT); }
+
{VarID} {
UnEscapeLexed(yytext+1);
llvmAsmlval.StrVal = strdup(yytext+1); // Skip %
- return VAR_ID;
+ return VAR_ID;
}
{Label} {
yytext[strlen(yytext)-1] = 0; // nuke colon
UnEscapeLexed(yytext);
- llvmAsmlval.StrVal = strdup(yytext);
- return LABELSTR;
+ llvmAsmlval.StrVal = strdup(yytext);
+ return LABELSTR;
+ }
+{QuoteLabel} {
+ yytext[strlen(yytext)-2] = 0; // nuke colon, end quote
+ UnEscapeLexed(yytext+1);
+ llvmAsmlval.StrVal = strdup(yytext+1);
+ return LABELSTR;
}
{StringConstant} { // Note that we cannot unescape a string constant here! The
- // string constant might contain a \00 which would not be
+ // string constant might contain a \00 which would not be
// understood by the string stuff. It is valid to make a
// [sbyte] c"Hello World\00" constant, for example.
//
- yytext[strlen(yytext)-1] = 0; // nuke end quote
- llvmAsmlval.StrVal = strdup(yytext+1); // Nuke start quote
- return STRINGCONSTANT;
+ yytext[strlen(yytext)-1] = 0; // nuke end quote
+ llvmAsmlval.StrVal = strdup(yytext+1); // Nuke start quote
+ return STRINGCONSTANT;
}
{PInteger} { llvmAsmlval.UInt64Val = atoull(yytext); return EUINT64VAL; }
-{NInteger} {
+{NInteger} {
uint64_t Val = atoull(yytext+1);
- // +1: we have bigger negative range
- if (Val > (uint64_t)INT64_MAX+1)
- ThrowException("Constant too large for signed 64 bits!");
- llvmAsmlval.SInt64Val = -Val;
- return ESINT64VAL;
+ // +1: we have bigger negative range
+ if (Val > (uint64_t)INT64_MAX+1)
+ ThrowException("Constant too large for signed 64 bits!");
+ llvmAsmlval.SInt64Val = -Val;
+ return ESINT64VAL;
}
{HexIntConstant} {
- llvmAsmlval.UInt64Val = HexIntToVal(yytext+3);
+ llvmAsmlval.UInt64Val = HexIntToVal(yytext+3);
return yytext[0] == 's' ? ESINT64VAL : EUINT64VAL;
}
-{EPInteger} { llvmAsmlval.UIntVal = atoull(yytext+1); return UINTVAL; }
+{EPInteger} {
+ uint64_t Val = atoull(yytext+1);
+ if ((unsigned)Val != Val)
+ ThrowException("Invalid value number (too large)!");
+ llvmAsmlval.UIntVal = unsigned(Val);
+ return UINTVAL;
+ }
{ENInteger} {
uint64_t Val = atoull(yytext+2);
- // +1: we have bigger negative range
- if (Val > (uint64_t)INT32_MAX+1)
- ThrowException("Constant too large for signed 32 bits!");
- llvmAsmlval.SIntVal = -Val;
- return SINTVAL;
+ // +1: we have bigger negative range
+ if (Val > (uint64_t)INT32_MAX+1)
+ ThrowException("Constant too large for signed 32 bits!");
+ llvmAsmlval.SIntVal = (int)-Val;
+ return SINTVAL;
}
{FPConstant} { llvmAsmlval.FPVal = atof(yytext); return FPVAL; }
{HexFPConstant} { llvmAsmlval.FPVal = HexToFP(yytext); return FPVAL; }
-[ \t\n] { /* Ignore whitespace */ }
+<<EOF>> {
+ /* Make sure to free the internal buffers for flex when we are
+ * done reading our input!
+ */
+ yy_delete_buffer(YY_CURRENT_BUFFER);
+ return EOF;
+ }
+
+[ \r\t\n] { /* Ignore whitespace */ }
. { return yytext[0]; }
%%
+